Triton-distributed-ascend
Language
English
中文
快速入门
安装
从源码构建
快速开始
发布说明
架构
分布式架构设计
API 参考
Triton-distributed 语义
triton_dist.language API
SHMEM 设备端 API
SHMEM 主机端 API
Swizzle 辅助接口
自动调优 API
开发者指南
Kernel 调试
环境变量与编译选项
多卡通算融合算子精度校验
单算子开发
性能优化指南
算子性能测试与调优:autotune 特性使用介绍
Unified Ascend operator tests
教程
全聚合-通用矩阵乘法(Allgather Gemm)
autotune 样例:UDMA Reverse All2All 增量优化实践
Reverse All2All (EP MoE Distributed Kernel)
常见问题
FAQ 文档
Triton-distributed-ascend
索引
索引
A
|
B
|
C
|
E
|
F
|
G
|
I
|
L
|
M
|
N
|
O
|
P
|
Q
|
R
|
S
|
T
|
W
A
aclshmem_align()
built-in function
aclshmem_calloc()
built-in function
aclshmem_create_tensor()
built-in function
aclshmem_finalize()
built-in function
aclshmem_free()
built-in function
aclshmem_free_tensor()
built-in function
aclshmem_get_unique_id()
built-in function
aclshmem_getmem()
built-in function
aclshmem_getmem_nbi()
built-in function
aclshmem_global_exit()
built-in function
aclshmem_info_get_name()
built-in function
aclshmem_info_get_version()
built-in function
aclshmem_init()
built-in function
aclshmem_init_using_unique_id()
built-in function
aclshmem_malloc()
built-in function
aclshmem_ptr()
built-in function
aclshmem_putmem()
built-in function
aclshmem_putmem_nbi()
built-in function
aclshmem_signal_wait_until()
built-in function
ACLSHMEMCmpOp(内置类)
ACLSHMEMSignalOp(内置类)
ACLSHMEMTeam(内置类)
aclshmemx_get_heap_base()
built-in function
aclshmemx_putmem_signal()
built-in function
aclshmemx_putmem_signal_nbi()
built-in function
ADD(ACLSHMEMSignalOp 属性)
B
barrier()
built-in function
barrier_all()
built-in function
barrier_all_vec()
built-in function
barrier_vec()
built-in function
built-in function
aclshmem_align()
aclshmem_calloc()
aclshmem_create_tensor()
aclshmem_finalize()
aclshmem_free()
aclshmem_free_tensor()
aclshmem_get_unique_id()
aclshmem_getmem()
aclshmem_getmem_nbi()
aclshmem_global_exit()
aclshmem_info_get_name()
aclshmem_info_get_version()
aclshmem_init()
aclshmem_init_using_unique_id()
aclshmem_malloc()
aclshmem_ptr()
aclshmem_putmem()
aclshmem_putmem_nbi()
aclshmem_signal_wait_until()
aclshmemx_get_heap_base()
aclshmemx_putmem_signal()
aclshmemx_putmem_signal_nbi()
barrier()
barrier_all()
barrier_all_vec()
barrier_vec()
consume_token()
extern_call()
fence()
getmem()
getmem_nbi()
int_p()
my_pe()
n_pes()
notify()
num_ranks()
pe_count()
putmem()
putmem_nbi()
putmem_signal()
putmem_signal_nbi()
quiet()
rank()
remote_ptr()
set_log_level()
signal_op()
signal_wait_until()
symm_at()
team_destroy()
team_my_pe()
team_n_pes()
team_split_2d()
team_split_strided()
team_translate_pe()
wait()
C
consume_token()
built-in function
E
EQ(ACLSHMEMCmpOp 属性)
extern_call()
built-in function
F
fence()
built-in function
G
getmem()
built-in function
getmem_nbi()
built-in function
GE(ACLSHMEMCmpOp 属性)
GT(ACLSHMEMCmpOp 属性)
I
InitAttr(内置类)
InitStatus(内置类)
int_p()
built-in function
INVALID(ACLSHMEMTeam 属性)
L
LE(ACLSHMEMCmpOp 属性)
LT(ACLSHMEMCmpOp 属性)
M
mem_size(InitAttr 属性)
my_pe()
built-in function
N
n_pes()
built-in function
NE(ACLSHMEMCmpOp 属性)
notify()
built-in function
num_ranks()
built-in function
O
OpEngineType(内置类)
P
pe_count()
built-in function
putmem()
built-in function
putmem_nbi()
built-in function
putmem_signal()
built-in function
putmem_signal_nbi()
built-in function
Q
quiet()
built-in function
R
rank()
built-in function
remote_ptr()
built-in function
S
set_log_level()
built-in function
SET(ACLSHMEMSignalOp 属性)
signal_op()
built-in function
signal_wait_until()
built-in function
symm_at()
built-in function
T
team_destroy()
built-in function
team_my_pe()
built-in function
team_n_pes()
built-in function
team_split_2d()
built-in function
team_split_strided()
built-in function
team_translate_pe()
built-in function
W
wait()
built-in function
WORLD(ACLSHMEMTeam 属性)